J/AJ/169/59 Kepler stars predicted rotation period: Deep Learning (Kamai+, 2025)
Accurate and robust stellar rotation periods catalog for 82771 Kepler stars
using deep learning.
Kamai I., Perets H.B.
<Astron. J., 169, 59 (2025)>
=2025AJ....169...59K 2025AJ....169...59K
ADC_Keywords: Stars, double and multiple ; Optical; Photometry; Models
Keywords: Light curves
Abstract:
We propose a new framework to predict stellar properties from light
curves. We analyze the light-curve data from the Kepler space mission
and develop a novel tool for deriving the stellar rotation periods for
main-sequence stars. Using this tool, we provide rotation periods for
more than 80K stars. Our model, LightPred, is a novel deep-learning
model designed to extract stellar rotation periods from light curves.
The model utilizes a dual-branch architecture combining long
short-term memory and transformer components to capture temporal and
global data features. We train LightPred on self-supervised
contrastive pretraining and simulated light curves generated using a
realistic spot model. Our evaluation demonstrates that LightPred
outperforms classical methods like the autocorrelation function in
terms of accuracy and average error. We apply LightPred to the Kepler
data set, generating the largest catalog to date. Using error analysis
based on learned confidence and consistency metric, we were able to
filter the predictions and remove stellar types with variability,
which is different than spot-induced variability. Our analysis shows
strong correlations between error levels and stellar parameters.
Additionally, we confirm tidal synchronization in eclipsing binaries
with orbital periods shorter than 10 days. Our findings highlight the
potential of deep learning in extracting fundamental stellar
properties from light curves, opening new avenues for understanding
stellar evolution and population demographics.
Description:
We downloaded the long cadence Kepler data set Data Release 25
(STScI 2016). The data is available at 10.17909/T9488N:
http://archive.stsci.edu/doi/resolve/resolve.html?doi=10.17909/T9488N
The Kepler spacecraft acquired data in the optical on a
115-square-degree region of the sky. The data were acquired on 29.4min
intervals, colloquially known as "long cadences". Long-cadence pixel
values were obtained by accumulating 270 consecutive 6.02s exposures.
Science acquisition of Q1 data began on 2009-05-13, and acquisition of
Q17 data concluded on 2013-05-11. This time period contains 71427
long-cadence intervals. A total of 198709 targets observed by Kepler
were searched for evidence of transiting planets in the final Q1-Q17
pipeline run.
Similarly to McQuillan+2014 (J/ApJS/211/24) and
Reinhold+2023 (J/A+A/678/A24), we omitted Q0-Q2 and Q17 and used the
data corrected for instrumental systematic using PDC-MAP
(Stumpe+, 2012PASP..124..985S 2012PASP..124..985S; Smith+, 2012PASP..124.1000S 2012PASP..124.1000S).
File Summary:
--------------------------------------------------------------------------------
FileName Lrecl Records Explanations
--------------------------------------------------------------------------------
ReadMe 80 . This file
table8.dat 35 82771 Final catalog of predicted periods
--------------------------------------------------------------------------------
See also:
V/133 : Kepler Input Catalog (Kepler Mission Team, 2009)
J/ApJ/687/1264 : Age estimation for solar-type dwarfs (Mamajek+, 2008)
J/A+A/534/A125 : Variability of A- and F-stars from Kepler (Uytterhoeven+ 2011)
J/ApJ/775/L11 : Stellar rotation periods for KOIs (McQuillan+, 2013)
J/A+A/557/L10 : Rotation periods of 12000 Kepler stars (Nielsen+, 2013)
J/A+A/560/A4 : Rotation periods of active Kepler stars (Reinhold+, 2013)
J/A+A/572/A34 : Pulsating solar-like stars in Kepler (Garcia+, 2014)
J/A+A/572/A34 : Pulsating solar-like stars in Kepler (Garcia+, 2014)
J/ApJS/211/24 : Rotation periods of Kepler MS stars (McQuillan+, 2014)
J/AJ/149/68 : A-F type variable stars from Kepler (Bradley+, 2015)
J/ApJ/801/3 : Rotation periods for Q3-Q14 KOIs (Mazeh+, 2015)
J/AJ/151/101 : Kepler Mission. VIII. 285 false positives (Abdul-Masih+, 2016)
J/AJ/152/158 : Final Kepler transiting planet search (DR25) (Twicken+, 2016)
J/ApJS/224/12 : Kepler planetary candidates. VII. 48-month (Coughlin+, 2016)
J/AJ/151/68 : Kepler Mission. VII. Eclipsing binaries in DR3 (Kirk+, 2016)
J/AJ/154/250 : Kepler EB classifications and rotation periods (Lurie+, 2017)
J/ApJS/229/30 : Revised stell. propert. of Q1-17 Kepler targets (Mathur+, 2017)
J/A+A/603/A52 : Activity cycles in 3203 Kepler stars (Reinhold+, 2017)
J/AJ/153/66 : Robo-AO Kepler Planetary Candidate Survey. III. (Ziegler+, 2017)
J/MNRAS/474/2094 : Inferring probabilistic stellar rot. periods (Angus+, 2018)
J/MNRAS/485/2380 : Gaia-derived lum. of Kepler A/F stars (Murphy+, 2019)
J/ApJS/244/21 : Surface rotation & activity of Kepler stars. I. (Santos+, 2019)
J/ApJ/871/174 : Kepler rapid rotators and Ks-band excesses (Simonian+, 2019)
J/AJ/159/280 : Gaia-Kepler stellar properties cat. I. KIC stars (Berger+, 2020)
J/ApJ/893/67 : Smoothed amplitudes from Kepler, K2 & TESS phot. (Morris, 2020)
J/ApJS/255/17 : Surface rot. & activity for Kepler stars. II. (Santos+, 2021)
J/MNRAS/510/3449 : KOI planets and KIC stars from Kepler DR25 (Bashi+, 2022)
J/A+A/678/A24 : New rotation periods of 67163 Kepler stars (Reinhold+, 2023)
J/A+A/689/A250 : Kepler DR25 cat. transit & detec. prob. (Castro-Gonzale+,2024)
http://archive.stsci.edu/doi/resolve/resolve.html?doi=10.17909/T9488N : Kepler
extracted light curves, all Quarters, long cadence only
Byte-by-byte Description of file: table8.dat
--------------------------------------------------------------------------------
Bytes Format Units Label Explanations
--------------------------------------------------------------------------------
1- 8 I8 --- KIC [757450/12984214] Kepler Input Catalog (V/133)
identifier
10- 15 F6.3 d Per [1.8/67.3] Predicted period
17- 21 F5.3 % ObsErr [4e-3/1.62] Observational error (1)
23- 27 F5.3 --- Conf [0.8/0.97] Model confidence (2)
29- 33 F5.3 d TotErr [4e-3/1.83] Total error (3)
35 I1 --- Flag [1/3]? Additional information on the sample (4)
--------------------------------------------------------------------------------
Note (1): Since each Kepler light-curve sample is potentially longer than eight
quarters, we can divide the full light curve into eight-quarter
segments and check the consistency of the model with respect to
different segments (for example, the consistency between predictions
for Q3-Q10, Q4-Q11, etc.). Since we have seven segments, each
corresponding to 720 days, there are in total 21 quarter pairs for
each sample. We find that the differences behave similarly to a
Normal distribution. We therefore construct the following
observational error: for each sample, we fit a normal distribution
using all of the pairs' differences, and take the 1σ value as
the observational error. See Sections 2.2, 2.3, and 5.2.1.
Note (2): See Section 4.2 for the model confidence equation.
Note (3): TotErr = ObsErr / ( Per (1 - Conf) ). See Section 5.2.3.
Note (4): Flag as follows:
1 = confirmed or candidate planet host star (1958 occurrences)
2 = Eclipsing Binary (399 occurrences)
3 = samples where ACF algorithm failed to predict period on one or more
segments (47051 occurrences)
--------------------------------------------------------------------------------
History:
From electronic version of the journal
(End) Prepared by [AAS], Robin Leichtnam [CDS] 18-Nov-2025